The QMS market is booming. The quality system isn't for sale.

Here is an uncomfortable observation – the fastest-growing product in quality right now is software, and none of it can close a nonconformance. Market analysts report a surge in QMS platform spending as manufacturers rush to digitise quality under cost pressure. I believe the numbers. I have closed nonconformances at midnight with a shift leader, a forklift driver and a whiteboard marker running dry. Nothing in that decision was ever sold per seat. The boom is real. What the market data cannot show is the gap between a licence and a system.

What the licence buys, and what it never will

A QMS platform is an evidence system. Records, routings, dashboards, audit trails, document control with actual revision discipline – all genuinely useful, all long overdue on plenty of shop floors I have walked. What no tier of the price list contains is decision rights. The subscription cannot say who contains a suspect batch at 22:40 on a Friday, who decides scrap versus sort, who signs the deviation, who owns the customer escape. That architecture lives in people and process, or it does not live at all.

I built the QA/QC department at SNOP, a 900-plus-employee greenfield plant, from the ground up. For the first long stretch it ran on whiteboards, QRQC in the aisles and ownership you could not hide from: one board, one owner per escape, one deadline per action, reviewed at fixed hours whether or not anyone felt like it. That department delivered a 70% reduction in defect cost and 98% customer satisfaction before any platform went live – not because anyone was nostalgic for paper, but because the reduction came from process design. Fast containment. Named owners. Disciplined problem-solving. Earlier, at WITTE Automotive, we took substantial cost out of failure the same way: QRQC, A3, and a Q-Wall that made status public and ownership personal. A login was never the constraint.

You cannot subscribe your way out of a decision you are afraid to make.

The configuration trap

Here is where the surge turns risky. Implementations are landing on shop floors this quarter, and too many are being configured by consultants who have never chaired an 8D, never run a containment decision, never had a customer quality engineer on the phone demanding answers before lunch. So the software does what software does. It encodes defaults. Every nonconformance routes to the quality manager regardless of failure mode. Root cause becomes a dropdown with "operator error" near the top. CAPA due dates are 30 days because 30 was in the demo environment.

Once live, the trap closes. Changing a routing becomes a change-management project – tickets, approvals, a validation cycle the vendor bills for. The platform makes a mediocre process non-negotiable. I have sat through IATF 16949 and AS9100 audits where a supplier projected the vendor's default workflow diagram and presented it as their quality system. The auditor asked one question: who decides containment on the night shift? The screen had no field for the answer. From there the audit drifted – the auditors interrogated configurations, the plant defended the tool, and nobody defended the outcome.

The consulting vantage confirms the pattern. Through FOREAST, the lean and quality consultancy I founded, I keep watching manufacturers – ArcelorMittal among the clients – buy the platform first and discover their decision map afterwards. The AI-inspection hype compounds it: smart cameras promise to accelerate the checking, which is welcome, but checking was never the bottleneck. Deciding was.

Buying without fooling yourself

None of this argues against digitising quality; I have argued for it in boardrooms myself. It argues for sequence. Before signature, map the nonconformance decision rights: who contains, who decides, who signs, at what € threshold – on nights and weekends as well as Tuesday mornings. If that map does not exist on paper, the platform will not invent it. It will photograph its absence.

Then pilot on one failure-cost stream measured in euros. Pick a recurring failure costing, say, €35,000 a month in sorting, rework and stopped lines; baseline time-to-containment and time-to-signature; let the business case be arithmetic rather than enthusiasm. Insist the platform accelerates QRQC and A3 rather than replacing them – the 24-hour discipline and the A3 thinking stay human, and the software's job is to make data arrive faster than the whiteboard did. And put defect owners in the configuration room, not just IT and the vendor's implementation team. The person who chases containment at 02:00 defines the fields and routings, or you get the demo workflow with your logo on the login page.

The spending surge is good news in exactly one case: buyers who know what they are buying. Software amplifies whatever system it lands on, including the absence of one. A licence with no decision architecture behind it is shelfware with an audit trail. When the escape reaches the customer, the auditor's question is unchanged – who owned it? Settle that before signature. The subscription renews annually; a customer escape only has to happen once.